OpenCV geometric shape recognition based on SVM

Many times, after Findcontours, to get a closed geometry, if known to be a circle, then think of ways to parameterize the expression of the circle, if it is known as a rectangle, the corresponding way to parameterize the expression of the rectangle. When we develop a simple linear or circular parametric expression algorithm, we will find out how to know whether it is a circle or a straight line, and become an urgent problem to be solved.

This paper uses SVM to realize geometric shape recognition. Currently only supports cirlce Rectangle triangle cross four species. Knowing the principles and ideas behind it and wanting to support more of the geometry's recognition, I believe, is not difficult.

Not much to say, the first part of the code.

1. Data to be trained

void GetData ()
{
	traindata.create (train_samples*classes, Samplesize.width*samplesize.height, CV_32FC1);
	Trainclasses.create (train_samples*classes, 1, CV_32FC1);
	Mat Src_image;
	Char file[255];
	int I, J;
	for (i = 0; i<classes; i++)
	{
		for (j = 0; j< train_samples; j + +)
		{
			sprintf (file, "Samples/s%d/%d.pn G ", I, j);
			Src_image = imread (file, 0);
			if (Src_image.empty ())
			{
				printf ("Error:cant load Image%s\n", file);
				Exit ( -1);
			}
			Mat image = Readimagesavecontour (src_image);
			Mat imagenewsize;
			Resize (image, Imagenewsize, samplesize, cv_inter_linear);
			Image.release ();
			Image = Imagenewsize.reshape (1, 1);
			Image.convertto (Traindata (range (I*train_samples + j, I*train_samples + j + 1), range (0, traindata.cols)), CV_32FC1); 
  trainclasses.at<float> (I*train_samples + j, 0) = i;
		}
	}
}

2. Affirm a SVM class and train

Cvsvmparams Svm_params;  The CVSVMPARAMS structure is used to define the basic parameters  
	Svm_params.svm_type = cvsvm::c_svc;     SVM type  
	svm_params.kernel_type = cvsvm::linear;//Do not map  
	svm_params.degree = 0;
	Svm_params.gamma = 1;
	SVM_PARAMS.COEF0 = 0;
	Svm_params. C = 1;
	svm_params.nu = 0;
	SVM_PARAMS.P = 0;
	Svm_params.term_crit = Cvtermcriteria (Cv_termcrit_iter, 0.01);

	CVSVM SVM;
	Svm.train (Traindata, Trainclasses, Mat (), Mat (), svm_params);

Well, that's the part of the code that's shown.

The download of the entire project and training sample is given below. Vs2013+opencv2.4.13 can be run directly. Download link

The effect is as shown in the figure